The invention is directed towards telecommunication and data communication applications. In particular, the invention is directed towards providing a fast switchover protection of communication outputs in these applications.
A communication system is a network of central offices connected by high-speed fiber optic lines. A Cesium (Cs) clock or global positioning system (GPS) signal at each central office synchronizes the communications. Communications networks further branch from the central offices and use lower speed connections. A building integration timing system (BITS) or BITS clock also called a network synchronization supply unit (SSU), is an electronics box at the central office that produces a timing or synchronization output signal connected to all the other network transmission and switching equipment in the office. A SSU contains a series of plug-in electronics cards including: an information management card (IMC), two input track and hold (ITH) cards or clock cards, one clock card for redundancy, with input from a Cs clock or GPS, and a series of output cards. The IMC, clock cards, and output cards connect to a backplane in the SSU. The output cards receive timing signals from the clock cards. The output cards have a power supply, electronics to communicate to the clock cards via the backplane, and a driver circuit that drives a transformer to produce the synchronization output signal. The output cards can have different types of synchronization output signals, e.g. DS0 or composite clock (CC), DS1 or T1, RS-422, and E1. DS0 operates at 64 Kb/sec. DS1 operates at 1.544 Mb/sec. RS-422 operates at various frequencies. E1 operates at 2.048 Mb/sec. DS1 and DS0 are common output types in the United States. E1 is common in Europe. These precise synchronization output signals are received by the other network elements, such as add/drop multiplexers (ADM), which carry the actual transmitted signals or data, called traffic. The synchronization output signals ensure that all out going transmissions from the central office have the same average frequency as the rest of the network.
It is important in data and telecommunication systems that synchronization output signals continue without loss in the event of a failure of an output card. Failure of an output card includes not only the actual failure of the card but also the accidental removal of a functioning card by a user. Replacement of output cards while the SSU is operating is called xe2x80x98hot swappingxe2x80x99.
N:1 is a prior art method of protecting outputs. There is one xe2x80x98hotxe2x80x99 spare output card for each type of output that can be switched in by a multiplexer if a failure occurs in one of the N output cards. Most slots in the SSU are filled with cards that have an output so there is a high utilization of outputs. For example, if the SSU has twenty slots for output cards and the MMC and two xe2x80x98hotxe2x80x99 spares occupy three, then there is an output utilization of 85%. Power efficiency is also about 85% since three out of the twenty cards consume power without producing an output. The switching is very slow because the multiplexer is an array of mechanical switches or relays. The switching time for the relays within the multiplexer is typically a few milliseconds. E1 output waveforms have a period of around 500 ns. This switching time results in the loss of several periods of output waveform. This is undesirable for most applications. The number of usable outputs is limited because a xe2x80x98hotxe2x80x99 spare must be kept for each output type. The switching increases in complexity for more than two output types. Only one output card is protected at a time, if another output card of the same output type fails before the first can be xe2x80x98hot swappedxe2x80x99, then that output signal is lost.
1:1 is another prior art method of output protection. There is one spare output card in a standby state for each output card. Half of the output cards in the SSU are not producing an output so there is only a 50% output utilization. Only the cards with outputs consume power so that the power efficiency is nearly 100%. The 1:1 arrangement provides fast switching and each output card is protected. The switching speed can be as fast as 500 ns or only one period of an E1 waveform.
1+1 is a third prior art method of output protection. This method is similar to the 1:1 except that the spare is xe2x80x98hotxe2x80x99. The 1+1 method has a 50% output utilization. Since all the output cards are consuming power and only half are producing outputs, the power efficiency is around 50%. The switching is very fast and each output card is protected. It is possible for the switch over to be nearly instantaneous if the output card fails in a open state. If the output card fails in the closed or shorted state, then the output card will have to be isolated and this can take 500 ns.
Prior art output cards have transformers generating heat and increasing the temperature of the output card. This can lead to accelerated failures of the more temperature sensitive components. The transformers themselves are very robust and rarely fail, yet they are replaced along with the rest of the output card when it fails. This is wasteful and costly.
The invention is a communication system having a series of output modules, each having a continuous synchronization output signal. Each output module has a driver assembly that has a driver output and a backup selector output that is capable of driving a failed next adjacent driver assembly. The driver assemblies have a power supply, microprocessor, field programmable gate array (FPGA), driver circuits and backup driver circuits. A sensor detects the failure of the driver or a backup driver output and generates a select signal in the failure state. A selector receives the driver output, the backup driver output, the select signal, and a select signal from the next adjacent driver assembly. The selector normally selects the driver output, but in response to a failure state from the select signal, the selector isolates the failed driver output. The previous adjacent driver assembly transmits a backup selector output to the failed driver assembly to maintain a continuous driver output. The first and the last output modules may be connected together so that each output is protected.
Protection switching is fast and each of the N output modules is protected. Output utilization is very high for this synchronization supply unit (SSU). The adjacent output module need not be of the same output type for the adjacent driver assembly to serve as protection, because the output module has a separate driver and backup driver. The output module is capable of simultaneously supplying two different output types. For example, a series of DS1 output modules can be adjacent to a series of DS0 output modules or the DS1 and DS0 output modules can even alternate. For a system with twenty output modules with any combination of output types, the SSU will have a 95% output utilization even if the first and last output modules of each series are not connected together. In this case, the one backup output module and all the backup drivers may be in the standby state for a power efficiency of nearly 100%. The backup drivers may be kept in standby or may be kept xe2x80x98hotxe2x80x99. The backup driver shares the power supply and most of the electronics, such as a microprocessor and field programmable gate array (FPGA) with the driver. As a result, if the backup driver is xe2x80x98hotxe2x80x99, power consumption may increase by only 30%. In this case, the backup output module and all the backup drivers may be kept xe2x80x98hotxe2x80x99 for a power efficiency of around 70%.
The driver assembly contains the active circuitry and a transformer card contains the transformers. Placing the transformers on a separate card reduces the thermal degradation of nearby electronics. In addition, transformers rarely fail so replacing them when a card fails is wasteful. This novel driver assembly can drive a variety of synchronization output signals, replacing several types of conventional output cards. In another preferred embodiment, the backup driver can be placed on separate backup driver card, so that a failed backup driver can be replaced without replacing the drivers. In this case, both driver and backup driver cards are stocked. The invention allows for xe2x80x98hot swappingxe2x80x99 of driver assemblies. This makes replacement of failed drivers or backup drivers easy.